Skip to content

count_prediction_metrics

Compute means, stds, iqr, entropy, jitter, label stability, and transform predictions to pd.Dataframe.

Return a 1D numpy array of predictions, 2D array of each model prediction for y_test, a data structure of metrics.

Parameters

  • y_true

    True labels

  • uq_results

    2D array of prediction proba for the zero value label by each model

  • with_predict_proba (bool) – defaults to True

    [Optional] A flag if model can return probabilities for its predictions. If no, only metrics based on labels (not labels and probabilities) will be computed.